hnij@bnl.UUCP ($user) (02/08/85)
Does anyone know of any RS232 interfaces for MIDI? I have a CP/M machine, and it has only one serial port. If no one knows of any, is there any interest in designing one? -- John Labovitz -- ..!decvax!philabs!sbcs!bnl!hnij
zadco@ssc-vax.UUCP (Rick Fairfield) (02/14/85)
> Does anyone know of any RS232 interfaces for MIDI? I have a CP/M machine, > and it has only one serial port. If no one knows of any, is there any > interest in designing one? > > -- > John Labovitz > -- > ..!decvax!philabs!sbcs!bnl!hnij MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, as you can see, MIDI is a problem over RS232. zzzzadco
bch@mcnc.UUCP (Byron Howes) (02/15/85)
In article <ssc-vax.416> zadco@ssc-vax.UUCP (Rick Fairfield) writes: >MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, >as you can see, MIDI is a problem over RS232. > zzzzadco I'm not sure I want to publicly disagree with my brother-in-law but... While the problem isn't trivial, It should be possible to send/translate RS232->MIDI in a fairly straightforward way. 32K bits/sec translates roughly to 3200 baud (as baud is used in lay terminology.) Assuming asyncronous protocol and loss of a bit to parity (why not) a 4800 baud link should be more than sufficient for the required MIDI bandwidth. -- Byron C. Howes ...!{decvax,akgua}!mcnc!ecsvax!bch
zadco@ssc-vax.UUCP (Rick Fairfield) (02/17/85)
> In article <ssc-vax.416> zadco@ssc-vax.UUCP (Rick Fairfield) writes: > > >MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, > >as you can see, MIDI is a problem over RS232. > > zzzzadco >> While the problem isn't trivial, It should be possible to send/translate >> RS232->MIDI in a fairly straightforward way. 32K bits/sec translates >> roughly to 3200 baud (as baud is used in lay terminology.) Assuming >> asyncronous protocol and loss of a bit to parity (why not) a 4800 baud >> link should be more than sufficient for the required MIDI bandwidth. >> -- >> Byron C. Howes >> ...!{decvax,akgua}!mcnc!ecsvax!bch Well. I'm confused on this issue. I thought buad translated roughly to bits/sec, so 9600 baud would be 9600 bits/sec. If what you say is true then someone should invent a MIDI-RS232 translator. It would have an immediate market; especially with owners of computers that don't have i/o slots, such as Kaypros. Any takers? zzzzzadco
hsu@cvl.UUCP (Dave Hsu) (02/18/85)
> In article <ssc-vax.416> zadco@ssc-vax.UUCP (Rick Fairfield) writes: > > >MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, > >as you can see, MIDI is a problem over RS232. > > zzzzadco > > While the problem isn't trivial, It should be possible to send/translate > RS232->MIDI in a fairly straightforward way. 32K bits/sec translates > roughly to 3200 baud (as baud is used in lay terminology.) Actually 32K bits/sec translates to 32K baud more or less. In the midi sense, it actually means 30.25(?)K baud. For this reason, most RS 232 ports cannot be run directly to midi because their baud-rate generators cannot handle this data rate. Most popular uarts (i.e. 6850 ACIA, AY5-1013) may be run at this speed, although it is near the maximum rating for some. In this way, the Apple Macintosh can be used to run midi devices, with little more than some opto-isolators for the current-loop interface. =Dave Hsu= "If only we'd had a cat like that..." (301) 454-4526 ARPA: hsu@cvl hsu@umd2 USnail: Computer Vision Laboratory CSNET: hsu@cvl Center for Automation Research BITNET: hsu@umd2 University of Maryland UUCP: {seismo,allegra,brl-bmd}!umcp-cs!cvl!hsu College Park, MD 20742
bch@mcnc.UUCP (Byron Howes) (02/19/85)
In article <ssc-vax.431> zadco@ssc-vax.UUCP (Rick Fairfield) writes: >Well. I'm confused on this issue. I thought buad translated roughly to >bits/sec, so 9600 baud would be 9600 bits/sec. > No. I'm the one that was confused. You are right, I must have been in never-never land when I posted that one. I shall go home, repent, and read my beginners computer glossary three times. -- Byron C. Howes ...!{decvax,akgua}!mcnc!ecsvax!bch
crandell@ut-sally.UUCP (Jim Crandell) (02/20/85)
> > In article <ssc-vax.416> zadco@ssc-vax.UUCP (Rick Fairfield) writes: > > > > >MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, > > >as you can see, MIDI is a problem over RS232. I really think someone should point out that 38.4 kb/s is a standard async serial speed, and it's generally not problematic over RS-232 unless (a) you're trying to drive five-mile-long cables with it or (b) the designer of the TTL/232 converter you're using got a little carried away trying to meet the specs for a class B (FCC) device rating so that the output slew rate is too low (don't laugh -- I've seen it done). Yes, it's true that the antiquated, PMOS AY5-1013 won't handle it, but the very common 8251 certainly will, and I believe I've read that the 6850 (or was it 6851?) has been used very effectively at 76.8 kb/s. So really, now, what's so bad about ~30 k? -- Jim Crandell, C. S. Dept., The University of Texas at Austin {ihnp4,seismo,ctvax}!ut-sally!crandell
dave@rocksvax.UUCP (02/20/85)
/***** rocksvax:net.music.synth / ssc-vax!zadco / 1:10 am Feb 20, 1985*/ > In article <ssc-vax.416> zadco@ssc-vax.UUCP (Rick Fairfield) writes: > > >MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, > >as you can see, MIDI is a problem over RS232. > > zzzzadco >Well. I'm confused on this issue. I thought buad translated roughly to >bits/sec, so 9600 baud would be 9600 bits/sec. RS232 has nothing to do with baud rates or data formatting, it is an electrical specification. It says things about what signal belongs on what pin, how the many volts these signals are, what level of electrical isolation required, and as a final note when sending data, how long a cable you can use for typical clock frequencies (effectively baud rates). Typical asynchronous formats use 10 bauds to transmit 8 bits. (1 start bit, 8 data, 1 stop bit) Assuming that the thing is async, I believe it is, then, 30000/8= 3750 bytes/sec which would be sent at a rate of 37500 baud. Most of the UART chips require a 16x clock to synchronize the data to its internal shift register so the chip requires a 600Khz clock. Most old UARTS just make this. I believe this was the reason they picked that rate and the crystal for the oscillator is an off the shelf value. Dave arpa: Sewhuk.HENR@Xerox.ARPA uucp: {allegra,rochester,amd,sunybcs}!rocksvax!dave
bek@duke.UUCP (Barrett E. Koster) (02/20/85)
Byron Howes and Rick Fairfield have been discussing 9600 baud vs. 9600 bps. I'm no hardware buff, but I think everybody's confused. (I believe) "baud" is symbols per second. "bps" is bits per second. If the symbols are 0 and 1, baud=bps. If, as is easily the case with multi-tone modems, there are more than two symbols, then several bits can be packed into one symbol. If there are N different tones, say, then X baud = log(N)X bps. What this has to do with the relevant reality in this case I have no idea. Barrett Koster duke!bek The thoughts expressed here are not necessarily thoughts.
graham@convex.UUCP (02/21/85)
Roland makes an RS-232 <-> MIDI interface (MMU-401 or something close to that). Look up a Roland ad in "Keyboard" magazine or call them @ (213)-685-5141. Marv Graham; Convex Computer Corp. {allegra,ihnp4,uiucdcs,ctvax}!convex!graham
brad@esc-bb.UUCP (Brad Benton) (02/21/85)
> If what you say is true then someone should invent a MIDI-RS232 translator. > It would have an immediate market; especially with owners of computers that > don't have i/o slots, such as Kaypros. Any takers? Sound Composition Systems has taken the challenge, and has announced the MIDI Performer II, an interface box that runs 9600 baud RS-232 on one side, and MIDI (31.25 Kbaud) on the other. All of this information comes from the latest (Vol 8, No 4) issue of Computer Music Journal. The box comes with a 2 Kbyte buffer, which can be expanded to 8K, and has a Z80 brain. And it works in two modes: transparent (bytes immediately available), and time tagged (you get to know exactly when a key was pressed, or you can tag info for a synth, and the MIDI box will schedule it for you). The "introductory price" is $345, and the place that makes and sells this device is: Sound Composition Systems 1277 East Colorado Pasadena, California 91106 (818) 440-0584 -bb ..!{seismo,ihnp4,gatech}!ut-sally!cyb-eng!esc-bb!brad
bart@reed.UUCP (Bart Massey) (02/22/85)
> In article <ssc-vax.416> zadco@ssc-vax.UUCP (Rick Fairfield) writes: > > >MIDI requires over 30000 bits/sec. RS232 usually runs at most 19200 baud so, > >as you can see, MIDI is a problem over RS232. > > zzzzadco > > I'm not sure I want to publicly disagree with my brother-in-law but... > > While the problem isn't trivial, It should be possible to send/translate > RS232->MIDI in a fairly straightforward way. 32K bits/sec translates > roughly to 3200 baud (as baud is used in lay terminology.) Assuming > asyncronous protocol and loss of a bit to parity (why not) a 4800 baud > link should be more than sufficient for the required MIDI bandwidth. > -- > > Byron C. Howes > ...!{decvax,akgua}!mcnc!ecsvax!bch Whoa! There's a decimal point problem here somewhere. 1 baud = 1 state change / second = 1 RS232 bit / second. 1K bits = 1024 bits , close to 1000 bits. SO 32K bits / second = 32K RS232 baud, close to *32000* baud. Comments? ..tektronix!reed!bart Massey